Enhance Responsive AI (Draft) to respect and align with the existing project-level Answer Library search settings, similar to how AI Assist currently functions. When AI is enabled at the project level, the system should consistently apply configured Answer Library controls (e.g., scoped collections, filtered sources, or restricted search settings) to AI-generated drafts.
Currently, AI Assist honors these project-level search constraints, but Responsive AI (Draft) does not, resulting in inconsistent AI behavior within the same project environment.
Use Case:
Enterprise customers (e.g., Oracle) configure project-level Answer Library search settings to ensure that only approved, relevant, or collection-specific content is used when generating responses. This is often required for:
- Governance and compliance controls
- Business unit–specific content restrictions
- Region- or product-specific answer scoping
- Maintaining accuracy and avoiding use of deprecated or restricted content
When Responsive AI (Draft) does not honor these settings, it may generate content outside the intended scope, reducing trust, increasing review time, and introducing compliance risk.
Aligning Responsive AI (Draft) with project-level Answer Library settings would provide:
- Consistent AI behavior across AI Assist and AI Draft
- Improved governance and predictability
- Reduced manual review and correction effort
- Greater enterprise confidence in AI-generated responses
